Finding Contours in OpenCV?

Question

When you retrieve contours from an image, you should get 2 contours per blob - one inner and one outer. Consider the circle below - since the circle is a line with a pixel width larger than one, you should be able to find two contours in the image - one from the inner part of the circle and one from the outer part.

Using OpenCV, I want to retrieve the INNER contours. However, when I use findContours (), I only seem to be getting the outer contours. How would I retrieve the inner contours of a blob using OpenCV?

I am using the C++ API, not C therefore only suggest functions that use the C++ API. (i.e. findContours () rather than cvFindContours ())

Thanks.

enter image description here

Well, I don't really have any unique code. I just have a Mat I retrieve from the camera. I run the canny edge detector on it, and then find contours using findContours (). All contours found are stored in a vector<vector<Point> >. Thats basically all I have so there isn't much point in me posting actual code - its just the traditional process of finding contours. — fdh
I have worked a bit with contours, though my interest was only the outer contours. I tried colororing a contour, with 'drawContours()'. Which exspects a sort of hierachy. The sample on 'drawContours()' that I used, did not color inner contours. What I'm trying to get at here, is that you probably need to use the hierachy array from 'findContours()' to sort them. (The reason I do not post an answer is that my knowledge of this hierachy is limited, an I would not want to confuse you on the matter. I have done a bit of testing on these and I could explain my understanding so far, if you request.) — Sonaten
Thanks for answering. If possible, I would appreciate it if you could explain your understanding of hierarchies so far. Thanks. — fdh
the external and internal boundary contours can be accessed using the h_next and v_next pointers in the CvSeq* structure where the contours are stored by the findContours() function. Here's a link from the book about how this is organised (I always dealt with just the external contours so there isn't any working code to post): books.google.co.in/… — AruniRC
I am using the C++ API not the C therefore I am not using CvSeq. I'm using vector<vector<Point> > — fdh

SSteve SSteve · Accepted Answer · 2011-12-11T20:15:56

I ran this code on your image and it returned an inner and outer contour.

#include <iostream>
#include "opencv2/core/core.hpp"
#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/highgui/highgui.hpp"

int main(int argc, const char * argv[]) {

    cv::Mat image= cv::imread("../../so8449378.jpg");
    if (!image.data) {
        std::cout << "Image file not found\n";
        return 1;
    }

    //Prepare the image for findContours
    cv::cvtColor(image, image, CV_BGR2GRAY);
    cv::threshold(image, image, 128, 255, CV_THRESH_BINARY);

    //Find the contours. Use the contourOutput Mat so the original image doesn't get overwritten
    std::vector<std::vector<cv::Point> > contours;
    cv::Mat contourOutput = image.clone();
    cv::findContours( contourOutput, contours, CV_RETR_LIST, CV_CHAIN_APPROX_NONE );

    //Draw the contours
    cv::Mat contourImage(image.size(), CV_8UC3, cv::Scalar(0,0,0));
    cv::Scalar colors[3];
    colors[0] = cv::Scalar(255, 0, 0);
    colors[1] = cv::Scalar(0, 255, 0);
    colors[2] = cv::Scalar(0, 0, 255);
    for (size_t idx = 0; idx < contours.size(); idx++) {
        cv::drawContours(contourImage, contours, idx, colors[idx % 3]);
    }

    cv::imshow("Input Image", image);
    cvMoveWindow("Input Image", 0, 0);
    cv::imshow("Contours", contourImage);
    cvMoveWindow("Contours", 200, 0);
    cv::waitKey(0);

    return 0;
}

Here are the contours it found:

findContour result image

Finding Contours in OpenCV?

3 Answers